| Input: |
| Part.1 |
| Image patches (positive and negative image patches), ; |
| Target(), ; |
| Output: |
| Tracking results (), at time ; |
| (1) |
| (2) if then |
| (3) Marker the target |
| (4) else |
| (5) Stage 1: using NCC and BF get the stable points. |
| (6) Stage 2: through Variance filter, Fern filter, NN filter get the best patches and staable point. |
| (7) Compare stage 1 with stage 2 get the best points. |
| (8) Update the Random Fren and positive, negative sample set. |
| (9) end if |
| Input: |
| Part.2 |
| Fish characteristic parameter: as data, ; |
| Model parameter: |
| include (base-score = 0.5, colsample-bylevel = 1, colsample-bytree = 1, gamma = 0, learning-rate = 0.1, max-delta-step = 0, max- |
| depth = 3, min-child-weight = 1, missing = None, -estimators = 100, thread = −1, objective = binary: logistic, reg-alpha = 0, |
| reg-lambda = 1, scale-pos-weight = 1, seed = 0, silent = True, subsample = 1) |
| Output: |
| Water Quality degree: ; model; Classification Accuracy |
| (10) Load data |
| (11) Split data into train and test sets by train-test-split( ). |
| (12) Load XGBClassifier and model.predict. |
| (13) Calculating Classification Accuracy |
